Erik on Software

Community Comment: Part 25

The comments I provided in reaction to a community discussion thread:
https://www.linkedin.com/feed/update/urn:li:activity:7001947850091675648?commentUrn=urn%3Ali%3Acomment%3A%28activity%3A7001947850091675648%2C7001953035828563968%29

Owner at Software Delivery Firm:

What does this say to you?

#agile #waterfall #softwaredevelopment #iterativedevelopment #incrementaldevelopment

Gfesser: The first two scenarios are optimistic: these assume (1) the product actually gets past the crap stage, and (2) the lifespan of the product ends up being long enough to see this happen (in contrast with starting from scratch again).

Media Query Source: Part 42

The responses I provided to a media outlet on December 5, 2022:

Media: I'm writing a post about financial advisors and AI. What is this technology? How does AI relate to areas like machine learning and deep learning? What is the importance of data? What are some things financial advisors need to know about this technology?

Gfesser: While incorrect, the terms "AI", "machine learning" and "deep learning" are often used interchangeably. AI ("artificial intelligence") is essentially about computing systems that can perform tasks typically associated with human intelligence. ML ("machine learning") is a subfield of AI which is all about task execution that relies on patterns rather than explicit instructions, and DL ("deep learning") is a type of ML based on artificial neural networks, the name of which hearkens to how the human brain filters information.

I personally don't care for the term "AI", and never use it in conversation, because it has been increasingly summoned by marketers who exaggeratingly pitch their products to what are often unsuspecting audiences, although exceptions do exist. For example, AWS ("Amazon Web Services") has acknowledged that the "AI services layer" it provides as part of its cloud offerings is really just a higher-level abstraction built on top of its ML services layer.

But data plays a crucial role regardless, because the value that AI, ML and DL provide is only as good as the data processed by it. Why is this the case? Because like any math equation, the output of a computing system's algorithms is based on the input. And if the input does not reflect "real" data, the output will likewise be unreliable. Perhaps you're familiar with the concept of GIGO ("garbage in, garbage out"). An ML algorithm or "model", for example, bases its predictions on input patterns, so if the data is problematic so will its predictions.

Financial advisors, as with other professional service providers, are expected to increasingly make use of AI over time to supplement (rather than replace) the services they provide their customers. While an advisor can generate portfolio strategies for customers without the use of AI, for example, the fact is that properly utilized AI can generate these relatively quickly with the appropriate input, additionally discovering patterns in the process that would likely otherwise go unnoticed.

---

See all of my responses to media queries here.